Prediction method for thickness of composite layer of composite board and related equipment

Abstract

The invention discloses a method for predicting the thickness of a composite layer of a composite board and related equipment, and relates to the technical field of composite board rolling, and the method comprises the following steps: obtaining assembly thickness data and target hot rolling process parameters of a to-be-measured composite board; inputting the assembly thickness data and the target hot rolling process parameters into a prediction model, and predicting the thickness of the composite plate to obtain a total thickness prediction value after cooling after hot rolling is finished; and based on the total thickness predicted value, calculating to obtain a composite layer thickness predicted value of the composite board. By fusing the assembly data, the hot rolling process parameters and the machine learning algorithm, online prediction and dynamic regulation and control of the thickness of the multi-layer are achieved.

Description

Technical Field

[0001] This application relates to the technical field of composite plate rolling, and particularly to a method for predicting the thickness of the clad layer of a composite plate and related equipment. Background Art

[0002] In the technical field of composite plate rolling, bimetallic composite plates such as stainless steel have extensive applications in many industries such as petroleum, chemical industry, papermaking, and metallurgy due to their combination of the characteristics of the clad layer and the base material. Composite plates are mostly produced by the double-billet stacking rolling mode. The clad layer is usually located in the core of the composite plate, making it difficult to directly monitor. During the production process, only the total thickness can be tracked online, and the thickness of the clad layer can only be obtained by measurement after slitting. Its uniformity cannot be corrected in a timely manner, which affects the quality and service performance of the product.

[0003] The prior art can only measure or monitor the total rolling thickness and cannot obtain the thickness of the clad layer of the composite plate online. The thickness of the clad layer needs to be measured manually with a thickness gauge after offline cutting. Therefore, there is an urgent need for a method for predicting the thickness of the clad layer of a composite plate to solve the above-mentioned problems. Summary of the Invention

[0004] A series of simplified concepts are introduced in the Summary of the Invention section, which will be further described in detail in the Detailed Description section. The Summary of the Invention section of this application does not mean to attempt to define the key features and essential technical features of the claimed technical solution, nor does it mean to attempt to determine the protection scope of the claimed technical solution.

[0005] In a first aspect, this application provides a method for predicting the thickness of the clad layer of a composite plate, including:

[0006] Obtain the billet thickness data of the composite plate to be measured and the target hot rolling process parameters. Among them, the billet thickness data includes the sampling point data of the plate head, plate middle, and plate tail, and the target hot rolling process parameters are obtained by correlation analysis and screening, including heating temperature, finishing rolling temperature, total reduction ratio, and rolling pass reduction distribution;

[0007] Input the billet thickness data and the target hot rolling process parameters into the prediction model to predict the thickness of the composite plate, so as to obtain the predicted value of the total thickness after cooling at the end of hot rolling. Among them, the prediction model is obtained through training for a preset number of times based on the historical billet thickness data and historical hot rolling process parameters of the composite plate;

[0008] Based on the predicted value of the total thickness, calculate the predicted value of the thickness of the clad layer of the composite plate.

[0009] In some embodiments, the sampling positions of the green board thickness data are within a first preset range from the width edges of the composite board, and at least 3 points are collected at the head, middle, and tail of each board number. Among them, the green board thickness data includes the thickness of the first upper cladding layer, the thickness of the first lower cladding layer, the thickness of the first upper composite board, and the thickness of the first lower composite board, and the first preset range is from 20 millimeters to 100 millimeters.

[0010] In some embodiments, the specific steps for screening target hot rolling process parameters through correlation analysis include:

[0011] Based on the Pearson coefficient, calculate the correlation coefficient between each candidate hot rolling process parameter and the actually measured total thickness value after hot rolling and cooling.

[0012] Select the candidate hot rolling process parameters whose absolute value of the correlation coefficient is greater than the preset threshold as the target hot rolling process parameters.

[0013] In some embodiments, the construction process of the prediction model includes:

[0014] Preprocess the historical green board thickness data and historical hot rolling process parameters to obtain preprocessed data.

[0015] Adopt the gradient boosting tree algorithm, use the preprocessed data as the training set, perform iterative training for a preset number of times. In each iteration, calculate the sample residuals, train a new weak learner for fitting, and calculate the weights of the weak learners to obtain the prediction model.

[0016] Use the genetic algorithm to optimize the prediction model to obtain the optimized prediction model.

[0017] In some embodiments, based on the predicted total thickness value, calculate the predicted cladding layer thickness value of the composite board, including:

[0018] According to the historical green board thickness data and historical hot rolling process parameters, determine the relationship coefficient between the actually measured total thickness after hot rolling and cooling and the total thickness of the upper and lower layers.

[0019] Based on the relationship coefficient and the predicted total thickness value, calculate the predicted total thickness values of the upper and lower layers after slitting.

[0020] According to the corresponding relationship between the base layer and the cladding layer, and the predicted total thickness values of the upper and lower layers after slitting, calculate the predicted cladding layer thickness value of the composite board.

[0021] In some embodiments, the method further includes:

[0022] Compare the predicted cladding layer thickness value with the preset tolerance range.

[0023] When the predicted value of the clad layer thickness is greater than the preset tolerance range, analyze the historical slab thickness data, historical hot rolling process parameters, current slab thickness data, and hot rolling process parameters to obtain deviation factor information;

[0024] Based on the deviation factor information, adjust the slab thickness data or hot rolling process parameters of the subsequent composite plate to be measured to optimize the prediction accuracy of the clad layer thickness of the composite plate.

[0025] In some embodiments, before inputting the slab thickness data and target hot rolling process parameters into the prediction model, it further includes:

[0026] Perform a normalization operation on the slab thickness data and target hot rolling process parameters.

[0027] In a second aspect, the present application proposes a prediction device for the clad layer thickness of a composite plate, including:

[0028] A data acquisition unit for acquiring the slab thickness data and target hot rolling process parameters of the composite plate to be measured, where the slab thickness data includes sampling point data at the head, middle, and tail of the plate, and the target hot rolling process parameters are obtained by correlation analysis and screening, including heating temperature, finishing rolling temperature, total reduction ratio, and rolling pass reduction distribution;

[0029] A model prediction unit for inputting the slab thickness data and target hot rolling process parameters into the prediction model to predict the thickness of the composite plate, so as to obtain the predicted value of the total thickness after cooling at the end of hot rolling, where the prediction model is obtained through training a preset number of times based on the historical slab thickness data and historical hot rolling process parameters of the composite plate;

[0030] A thickness calculation unit for calculating the predicted value of the clad layer thickness of the composite plate based on the predicted value of the total thickness.

[0031] In a third aspect, an electronic device includes: a memory, a processor, and a computer program stored in the above memory and executable on the processor, and the processor is used to implement the steps of the prediction method for the clad layer thickness of the composite plate according to any one of the first aspects when executing the computer program stored in the memory.

[0032] In a fourth aspect, the present application also proposes a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the prediction method for the clad layer thickness of the composite plate according to any one of the first aspects.

[0033] In summary, the present application obtains the data of the thickness of the assembled blank and the target hot rolling process parameters screened through correlation analysis, and uses a machine learning prediction model to achieve accurate prediction of the thickness of the clad layer of the composite plate. By adopting the gradient boosting tree algorithm and the genetic algorithm to optimize the model, the present invention can effectively establish the non-linear relationship between the thickness of the assembled blank, the process parameters and the thickness of the clad layer of the composite plate. Compared with the prior art, the present invention can not only realize the real-time prediction of the thickness of the composite plate, but also dynamically adjust the production process parameters, thereby improving the prediction accuracy and production efficiency, reducing the rejection rate, and providing an efficient and intelligent technical solution for industrial production. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] By reading the detailed description of the preferred embodiments below, various other advantages and benefits will become clear to those of ordinary skill in the art. The drawings are only for the purpose of illustrating the preferred embodiments and are not considered to be a limitation of this specification. Moreover, throughout the drawings, the same reference numerals are used to represent the same components. In the drawings:

[0035] Figure 1 It is a schematic flow chart of a method for predicting the thickness of the clad layer of a composite plate provided by an embodiment of the present application;

[0036] Figure 2 It is a schematic diagram of the sampling data of the assembled blank provided by an embodiment of the present application;

[0037] Figure 3 It is a schematic diagram of the sampling data of the finished composite plate provided by an embodiment of the present application;

[0038] Figure 4 It is a schematic diagram of the thickness data of the assembled blank provided by an embodiment of the present application;

[0039] Figure 5 It is a schematic diagram of the thickness data of the finished composite plate provided by an embodiment of the present application;

[0040] Figure 6 It is a schematic diagram of the structure of a device for predicting the thickness of the clad layer of a composite plate provided by an embodiment of the present application;

[0041] Figure 7 It is a schematic diagram of the structure of an electronic device for predicting the thickness of the clad layer of a composite plate provided by an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0042] In the description, claims, and above-mentioned drawings of this application, the terms "first", "second", "third", "fourth", etc. (if any) are used to distinguish similar objects and do not necessarily describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances so that the embodiments described here can be implemented in an order different from that illustrated or described here. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units does not necessarily limit to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products, or devices. Next, the technical solutions in the embodiments of this application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of this application. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all the embodiments.

[0043] Please refer to Figure 1 , which is a schematic flow chart of a method for predicting the thickness of the composite layer of a composite board provided by an embodiment of this application. Specifically, it may include:

[0044] S110. Obtain the blank thickness data of the composite board to be measured and the target hot rolling process parameters. Among them, the blank thickness data includes the sampling point data of the head, middle, and tail of the board. The target hot rolling process parameters are obtained through correlation analysis and include heating temperature, finishing rolling temperature, total reduction ratio, and reduction distribution per rolling pass;

[0045] Exemplarily, the core of step S110 is to obtain the blank thickness data of the composite board and the target hot rolling process parameters, and screen out the process parameters most relevant to the composite layer thickness through correlation analysis. The blank thickness data includes the sampling point data from different positions (head, middle, and tail) of the composite board. By measuring at the three key positions of the head, middle, and tail of the board respectively, the thickness changes in different regions of the composite board can be better captured, which provides comprehensive basic data for the subsequent prediction of the composite board thickness. The selection of these sampling points is of great significance for analyzing the thickness uniformity and regional differences of the composite board during the rolling process.

[0046] In addition, the screening of the target hot rolling process parameters is achieved through correlation analysis. By calculating the correlation coefficient between each process parameter and the total thickness of the composite plate, the process parameters most relevant to the change in the thickness of the composite plate are selected. Specifically, the heating temperature, finishing rolling temperature, total reduction ratio, and reduction distribution in each rolling pass are selected as the target process parameters. These parameters have a direct impact on the material flow, deformation, and final thickness during hot rolling, so there is a strong correlation between them and the total thickness and clad layer thickness of the composite plate. Through this screening method, it can be ensured that the selected process parameters can reflect the change law of the composite plate thickness to the greatest extent, thereby improving the accuracy and reliability of the subsequent prediction model.

[0047] S120. Input the billet thickness data and the target hot rolling process parameters into the prediction model to predict the thickness of the composite plate, so as to obtain the predicted value of the total thickness after cooling at the end of hot rolling, where the prediction model is obtained through training for a preset number of times based on the historical billet thickness data and historical hot rolling process parameters of the composite plate;

[0048] Exemplarily, in step S120, first input the previously collected billet thickness data and the screened target hot rolling process parameters into the prediction model to predict the thickness of the composite plate. The core objective of this step is to predict the total thickness after cooling at the end of hot rolling. In this process, the billet thickness data and process parameters are used as input features and processed through the established prediction model. This model needs to associate these input information with historical data to infer the final thickness of the composite plate after passing through the hot rolling process. To ensure the accuracy and reliability of the prediction results, the model is trained based on the historical data of the composite plate (including historical billet thickness data and historical hot rolling process parameters).

[0049] The prediction model is obtained by inputting historical data into the model and training for a preset number of times. During the training process, the model gradually learns how to establish the relationship between the input billet thickness data and the corresponding process parameters and the final thickness of the composite plate. In each iterative training, the model will adjust according to the previous error to continuously optimize its prediction ability. Through multiple iterations, the model can accurately capture complex non-linear relationships and can accurately predict the total thickness of the composite plate based on the input data. The effectiveness and accuracy of this prediction model depend on the richness and diversity of historical data, ensuring that its application in actual production can accurately reflect the impact of the hot rolling process on the thickness of the composite plate.

[0050] S130. Calculate the predicted value of the clad layer thickness of the composite plate based on the predicted value of the total thickness.

[0051] Exemplarily, in step S130, based on the predicted total thickness value obtained through the prediction model, the predicted clad layer thickness value of the composite plate is further calculated. The predicted total thickness value is the overall thickness of the composite plate after cooling at the end of hot rolling, and the clad layer thickness is a key component of the total thickness. Its calculation requires combining the formed upper and lower layer thickness relationship model of the slab thickness data and historical process parameters.

[0052] Specifically, based on the historical slab thickness data and hot rolling process parameters, first determine the proportional relationship between the total thickness and the total thickness of the upper and lower layers. Then, according to these relationship coefficients and the predicted total thickness value, calculate the predicted total thickness values of the upper and lower layers after slitting. Further, combining the thickness distribution law of the upper and lower layers and the corresponding relationship between the base layer and the clad layer, the predicted value of the clad layer thickness of the composite plate can be deduced. This calculation method effectively utilizes the correlation between the slab data and process parameters, and realizes the accurate prediction of the clad layer thickness by analyzing the contribution distribution of the total thickness. This step not only improves the accuracy of the clad layer thickness prediction, but also provides a direct basis for the subsequent production process adjustment.

[0053] In some examples, the sampling position of the slab thickness data is within the first preset range from the width edge of the composite plate, and at least 3 points are collected at the head, middle, and tail of each plate number respectively. Among them, the slab thickness data includes the first upper clad layer thickness, the first lower clad layer thickness, the first upper composite plate thickness, and the first lower composite plate thickness, and the first preset range is 20 millimeters to 100 millimeters.

[0054] Exemplarily, as Figure 2 shown, in the composite plate clad layer thickness prediction method, the sampling position and sampling method of the slab thickness data are crucial for the prediction accuracy. The sampling position of the slab thickness data is usually within the range of 20 millimeters to 100 millimeters from the width edge of the composite plate. This range is selected to avoid measurement deviations caused by uneven material flow in the edge area during the rolling process. At the same time, data of at least 3 points are collected at the head, middle, and tail of each plate number respectively to ensure that the distribution characteristics of the slab thickness in the length direction can be comprehensively reflected. This multi-point sampling strategy can effectively capture the spatial variation of the slab thickness and provide high-precision input data for the subsequent prediction model. As Figure 3 shown, it is a schematic diagram of the sampling data of the composite plate finished product. Similar to the sampling position of the slab thickness data, within the range of 1 meter to 2 meters from the width edge of the composite plate, data of at least 3 points are collected at the head, middle, and tail of each plate number respectively.

[0055] As Figure 3 shown, in the embodiment of the present application, taking the composite plate with the clad layer being a stainless steel layer and the base layer being a carbon steel layer as an example, the slab thickness data includes the first upper clad layer thickness t 1 , the first lower clad layer thickness t 2, the thickness t of the first upper composite plate 3 and the thickness t of the first lower composite plate 4 . Among them, the thickness of the first upper clad layer and the thickness of the first lower clad layer respectively represent the initial thickness of the stainless steel clad layer in the upper and lower layers of the composite plate, while the thickness of the first upper composite plate and the thickness of the first lower composite plate represent the total thickness of the clad layer and the base layer (such as carbon steel) in the upper and lower layers. These data jointly describe the initial state of the billet assembly stage and provide key input features for the prediction model. Through the multi-point and multi-parameter acquisition method, the distribution law of the billet assembly thickness can be more accurately reflected, thereby improving the reliability of the prediction model. Similarly, Figure 4 is a schematic diagram of the thickness data of the finished composite plate, including the thickness t of the second upper clad layer 5 , the thickness t of the second lower clad layer 6 , the thickness t of the second upper composite plate 7 and the thickness t of the second lower composite plate 8 , among which, the thickness t of the second upper clad layer 5 , the thickness t of the second lower clad layer 6 is the clad layer thickness value to be predicted in this application.

[0056] In some examples, the specific steps of screening target hot rolling process parameters through correlation analysis include:

[0057] Based on the Pearson coefficient, calculate the correlation coefficient between each candidate hot rolling process parameter and the total thickness value actually measured after hot rolling and cooling.

[0058] Select the candidate hot rolling process parameters whose absolute value of the correlation coefficient is greater than the preset threshold as the target hot rolling process parameters.

[0059] Exemplarily, first calculate the correlation coefficient between each candidate hot rolling process parameter and the total thickness value actually measured after hot rolling and cooling based on the Pearson coefficient. The Pearson coefficient is used to measure the linear correlation between two variables, and its value range is [-1, 1]. The closer the absolute value is to 1, the stronger the correlation. By calculating the correlation coefficient between each candidate parameter (such as heating temperature, finishing temperature, reduction ratio, etc.) and the total thickness, the influence degree of these parameters on the total thickness can be quantified.

[0060] After calculating the correlation coefficient, candidate hot rolling process parameters with absolute values greater than a preset threshold (such as 0.7) are selected as the target hot rolling process parameters. The purpose of this step is to eliminate parameters that have less impact on the total thickness, thereby reducing the model complexity and improving the prediction efficiency. For example, if the correlation coefficient between the heating temperature and the total thickness is 0.7, while the correlation coefficient of a certain pass reduction rate is 0.3, then the heating temperature will be retained as the target parameter, and this pass reduction rate may be eliminated. In this way, the selected target hot rolling process parameters can more accurately reflect the main influencing factors on the total thickness during the rolling process.

[0061] In the embodiments of this application, the selected target hot rolling process parameters include heating temperature, rolling pass reduction distribution, finishing temperature, intermediate slab thickness, total reduction rate, and pass reduction rate. The heating temperature and the finishing temperature directly affect the plasticity of the material, and the rolling pass reduction distribution and the pass reduction rate reflect the strain accumulation characteristics during the material deformation process. By inputting these strongly correlated parameters into the prediction model, the accuracy of the total thickness prediction can be significantly improved, providing a reliable basis for the subsequent calculation of the cladding thickness.

[0062] In some examples, the construction process of the prediction model includes:

[0063] Preprocess the historical slab thickness data and historical hot rolling process parameters to obtain preprocessed data;

[0064] Adopt the gradient boosting tree algorithm, use the preprocessed data as the training set, perform iterative training for a preset number of times, calculate the sample residuals in each iteration, train a new weak learner for fitting, and calculate the weights of the weak learners to obtain the prediction model;

[0065] Use the genetic algorithm to optimize the prediction model to obtain the optimized prediction model.

[0066] Exemplarily, preprocess the historical slab thickness data and historical hot rolling process parameters to ensure data quality and meet the requirements of model training. The preprocessing steps include eliminating null values, outliers, and redundant data, and at the same time normalizing the data to eliminate the dimensional differences between different parameters. The preprocessed data corresponds one by one with the hot rolling process parameters according to the plate number, forming a high-quality training set, laying a foundation for subsequent model training.

[0067] The GBDT (Gradient Boosting Decision Tree) algorithm is used for model training. GBDT is an ensemble learning algorithm that uses CART regression trees as weak learners and gradually optimizes the prediction results through multiple rounds of iteration. In each round of iteration, the model calculates the sample residuals (i.e., the difference between the actual value and the current predicted value) and trains a new weak learner to fit these residuals. Subsequently, the weights of each weak learner are calculated and added to the current model to gradually reduce the prediction error. Through a preset number of iterations, a strong learner, i.e., an optimized prediction model, is finally obtained. The advantage of GBDT is that it can automatically capture non-linear relationships and handle high-dimensional feature data, making it suitable for the complex scenario of composite board thickness prediction.

[0068] To further improve the prediction accuracy of the model, the GA (Genetic Algorithm) is used to optimize the hyperparameters of the GBDT model. The genetic algorithm dynamically adjusts the hyperparameters of GBDT (such as tree depth, learning rate, number of leaf nodes, etc.) by simulating the natural selection process to find the optimal parameter combination. The global search ability of the genetic algorithm can effectively prevent the model from falling into local optima, thereby improving the generalization ability and robustness of the prediction model. Finally, the optimized prediction model can accurately predict the total thickness after hot rolling and cooling based on the input slab thickness data and hot rolling process parameters, providing a reliable basis for calculating the cladding thickness.

[0069] In some instances, based on the predicted total thickness value, the predicted cladding thickness value of the composite board is calculated, including:

[0070] According to the historical slab thickness data and historical hot rolling process parameters, determine the relationship coefficient between the actually measured total thickness after hot rolling and cooling and the total thickness of the upper and lower layers;

[0071] Based on the relationship coefficient and the predicted total thickness value, calculate the predicted total thickness values of the upper and lower layers after slitting;

[0072] According to the corresponding relationship between the base layer and the cladding layer, and the predicted total thickness values of the upper and lower layers after slitting, calculate the predicted cladding thickness value of the composite board.

[0073] Exemplarily, in the production process of composite boards, the hot rolling process plays a key role in the change of composite board thickness. A large amount of information is contained in the historical data, reflecting the relationship between the parameters of each production link and the final thickness. The thickness of the second upper composite board is t 7 , and the thickness of the second lower composite board is t 8 , and the actually measured total thickness after hot rolling and cooling is T.

[0074] To establish the relationship between the total thickness and the thickness of the upper and lower layers, it is expressed as:

[0075] T = a·t 7 + b·t 8

[0076] Where a and b are the relationship coefficients to be determined.

[0077] Based on historical data, using mathematical methods such as regression analysis and the least squares method, the data is fitted and calculated to solve for the values of a and b that maximize the matching degree between the model and the actual data, thereby clarifying the quantitative relationship between the total thickness and the thicknesses of the upper and lower layers.

[0078] Based on the determined relationship coefficients and the predicted value of the total thickness, calculate the predicted values of the total thicknesses of the upper and lower layers after slitting. After obtaining the predicted value of the total thickness through the prediction model Let the predicted value of the thickness of the upper composite board after slitting be The predicted value of the thickness of the lower composite board be Based on some characteristics of the upper and lower layer materials during the rolling process, such as the deformation ratio k, etc., combined with the relationship coefficients and the predicted value of the total thickness, solve for the predicted values of the upper and lower layer thicknesses by solving a system of equations, expressed as:

[0079]

[0080] By solving this system of equations, obtain the predicted values of the thicknesses of the upper and lower composite boards after slitting and

[0081] According to the corresponding relationship between the base layer and the cladding layer, and the predicted values of the total thicknesses of the upper and lower layers after slitting, calculate the predicted value of the thickness of the cladding layer of the composite board (i.e., the predicted values of t 5 、t 6 ). The upper composite board consists of an upper cladding layer (with a thickness of t 5 ) and an upper base layer, and the lower composite board consists of a lower cladding layer (with a thickness of t 6 ) and a lower base layer.

[0082] The predicted value of the total thickness of the upper composite board is Then where ub is the predicted value of the thickness of the upper base layer; combining historical data and process knowledge, the thickness change law of the upper cladding layer and the upper base layer during the rolling process is The predicted value of the thickness of the upper cladding layer can be obtained The predicted value of the total thickness of the lower composite board is Then where lb is the predicted value of the thickness of the lower base layer; combining historical data and process knowledge, the thickness change law of the lower cladding layer and the lower base layer during the rolling process is The predicted value of the thickness of the lower cladding layer can be obtained

[0083] Through the above rigorous calculation process based on historical data and process relationships, the predicted value of the composite plate's multilayer thickness can be gradually and accurately derived starting from the predicted value of the total thickness; this method not only simplifies the calculation process of the multilayer thickness, but also ensures the accuracy and reliability of the prediction results through dynamic calibration of historical data, providing an important basis for the refined control of composite plate production.

[0084] In some examples, it also includes:

[0085] Compare the predicted layer thickness with the preset tolerance range;

[0086] When the predicted value of the composite layer thickness is greater than the preset tolerance range, historical batch thickness data, historical hot rolling process parameters, and current batch thickness data and hot rolling process parameters are analyzed to obtain deviation factor information;

[0087] Based on the deviation factor information, the subsequent assembly thickness data or hot rolling process parameters of the composite plate to be tested are adjusted to optimize the prediction accuracy of the composite plate layer thickness.

[0088] For example, in the composite plate composite layer thickness prediction method, the verification and feedback control of the prediction results are important links to ensure production quality. Compare the predicted value of the composite layer thickness with the preset tolerance range to determine whether the predicted value meets the production requirements. If the predicted value of the composite layer thickness exceeds the tolerance range, the cause of the deviation needs to be further analyzed. By comparing the historical assembly thickness data, historical hot rolling process parameters, and the current assembly thickness data and hot rolling process parameters, the key factors causing the deviation (such as uneven initial thickness of the assembly composite layer, unreasonable distribution of reduction rate, etc.) are identified, thereby obtaining deviation factor information.

[0089] Based on the deviation factor information, the subsequent assembly thickness data or hot rolling process parameters of the composite plate to be tested are adjusted to optimize the prediction accuracy of the composite layer thickness. For example, if the deviation analysis shows that the heating temperature is too high, resulting in thinning of the composite layer, the heating temperature in subsequent production can be appropriately reduced; if the initial thickness of the composite layer of the assembly is uneven, the assembly design can be adjusted to improve the thickness uniformity. Through this closed-loop feedback mechanism, the production process parameters can be dynamically optimized, the deviation of the composite layer thickness can be reduced, and the product qualification rate and yield rate can be improved.

[0090] It should be noted that, in the embodiment of the present application, the preset tolerance range is 1.0±0.1 mm.

[0091] In some examples, before inputting the batch thickness data and the target hot rolling process parameters into the prediction model, the method further includes:

[0092] The batch thickness data and target hot rolling process parameters are normalized.

[0093] Exemplarily, the blank thickness data and the target hot rolling process parameters often have different dimensions and value ranges. For example, the blank thickness data is usually in millimeters, and the value range may be between a few millimeters and dozens of millimeters; while the heating temperature, as one of the target hot rolling process parameters, is in degrees Celsius, and the value range may reach several hundred or even over a thousand degrees Celsius. Different dimensions and value ranges will cause unbalanced effects on the data during the model training process. Those parameters with larger value ranges may dominate in the model training, causing the model to overly focus on the changes of these parameters while ignoring other parameters with smaller but equally important value ranges. The core purpose of the normalization process is to eliminate this difference in dimension and value range, mapping all data to a unified standard range, such as the common [0,1] interval. Through the normalization process, each parameter can have the same weight status in the model training, avoiding the model training deviation caused by the difference in dimension and value range, so that the model can more fairly learn the relationship between each parameter and the composite plate thickness, improving the stability and accuracy of the model training.

[0094] In the embodiments of the present application, Table 1 lists the comparison results of the predicted total thickness data and the measured total thickness data under different steel grades and specifications.

[0095] Table 1 Data Comparison Results / mm

[0096] Example Steel grade Predicted total thickness mean value Measured total thickness mean value Predicted cladding thickness mean value Measured cladding thickness mean value Example 1 304+235B 6 5.98 1 0.97 Example 2 316+235B 8 8.02 2 2.01 Example 3 304L + 355B 6 6.03 1 1.03 Example 4 316L + 355B 10 10.1 2 2.05 Example 5 410+355B 5 4.99 1 0.98 Example 6 304L + 345R 40 40.3 4 4.11 Example 7 904L + 345R 50 50.4 3 3.14

[0097] It can be seen from the example data in Table 1 that the composite plate thickness prediction method of the present application shows high prediction accuracy under various steel grade combinations. Whether it is the predicted value of the total thickness or the cladding thickness, the error between its mean value and the actual measured value is controlled within a reasonable range. In Example 1, the predicted total thickness mean value is 6 mm, and the error from the measured value of 5.98 mm is only 0.02 mm, and the prediction error of the cladding thickness is 0.03 mm; in Example 7, even for the 904L + 345R composite plate with a relatively large total thickness, the prediction errors of its total thickness and cladding thickness are only 0.4 mm and 0.14 mm respectively. This shows that the prediction method can accurately estimate the total thickness and cladding thickness of the composite plate, providing a reliable reference basis for actual production, helping to effectively control the composite plate thickness during the production process, improving the product quality stability, and reducing quality problems caused by thickness deviation.

[0098] Please refer to Figure 6 , which is a schematic structural diagram of a device for predicting the cladding thickness of a composite plate provided by the embodiments of the present application, including:

[0099] A data acquisition unit 21 is configured to acquire the blank thickness data of the composite plate to be measured and the target hot rolling process parameters. The blank thickness data includes the sampling point data at the plate head, the plate middle, and the plate tail. The target hot rolling process parameters are obtained by correlation analysis and include the heating temperature, the finishing rolling temperature, the total reduction ratio, and the reduction distribution per rolling pass.

[0100] A model prediction unit 22 is configured to input the blank thickness data and the target hot rolling process parameters into a prediction model to predict the thickness of the composite plate, so as to obtain the predicted value of the total thickness after cooling at the end of hot rolling. The prediction model is obtained by training a preset number of times based on the historical blank thickness data and the historical hot rolling process parameters of the composite plate.

[0101] A thickness calculation unit 23 is configured to calculate the predicted value of the cladding thickness of the composite plate based on the predicted value of the total thickness.

[0102] Please refer to Figure 7 , an electronic device 300 is further provided in an embodiment of the present application, including a memory 310, a processor 320, and a computer program 311 stored on the memory 310 and executable on the processor. When the processor 320 executes the computer program 311, the steps of any method of the prediction device for the cladding thickness of the composite plate are implemented.

[0103] Since the electronic device introduced in this embodiment is the device adopted for implementing a prediction device for the cladding thickness of a composite plate in an embodiment of the present application, based on the method introduced in the embodiment of the present application, those skilled in the art can understand the specific implementation manners and various variations of the electronic device in this embodiment. Therefore, the specific implementation of how this electronic device implements the method in the embodiment of the present application will not be described in detail here. As long as the device adopted by those skilled in the art to implement the method in the embodiment of the present application belongs to the scope to be protected by the present application.

[0104] In the specific implementation process, when the computer program 311 is executed by the processor, any implementation manner in the corresponding embodiment of the first aspect can be implemented.

[0105] It should be noted that in the above embodiments, the descriptions of the various embodiments have their own emphases. For the parts not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0106] Those skilled in the art should understand that the embodiments of the present application may provide a method, a system, or a computer program product. Therefore, the present application may be implemented in the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application may be implemented in the form of a computer program product implemented on one or more computer-readable storage media (including but not limited to disk memories, CD-ROMs, optical memories, etc.) containing computer-readable program codes.

[0107] This application is described with reference to the flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the present application. It should be understood that each flow and / or block in the flowchart and / or block diagram can be implemented by computer program instructions, and the combination of flows and / or blocks in the flowchart and / or block diagram can also be implemented. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in one Figure 1 flow or multiple flows and / or blocks Figure 1 block or multiple blocks.

[0108] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including instruction means that implement the functions specified in one Figure 1 flow or multiple flows and / or blocks Figure 1 block or multiple blocks.

[0109] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one Figure 1 flow or multiple flows and / or blocks Figure 1 block or multiple blocks.

[0110] Embodiments of the present application also provide a computer program product that includes computer software instructions. When the computer software instructions run on a processing device, the processing device is caused to execute Figure 1 the process of a method for predicting the thickness of the composite layer of a composite board in a corresponding embodiment.

[0111] A computer program product includes one or more computer instructions. When the computer instructions are loaded and executed on a computer, they produce, wholly or partly, a process or a function in accordance with the embodiments of the present application. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or any other programmable device. The computer instructions may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another. For example, the computer instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center in a wired manner (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or a wireless manner (such as infrared, wireless, microwave, etc.). The computer-readable storage medium may be any available medium that can be stored by a computer or a data storage device such as a server or a data center that includes one or more available media integrated. The available medium may be a magnetic medium (such as a floppy disk, a hard disk, a magnetic tape), an optical medium (such as a DVD), or a semiconductor medium (such as a solid state disk (SSD)), etc.

[0112] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices, and units described above may refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0113] In several embodiments provided in the present application, it should be understood that the disclosed devices, apparatuses, and methods may be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and there may be other division methods in actual implementation. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Another point is that the couplings, direct couplings, or communication connections shown or discussed with each other may be indirect couplings or communication connections through some interfaces, devices, or units, and may be in electrical, mechanical, or other forms.

[0114] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0115] In addition, in each embodiment of the present application, each functional unit can be integrated into a processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of a software functional unit.

[0116] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods in the various embodiments of the present application. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs that can store program codes.

[0117] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the various embodiments of the present application.

[0118] Although the preferred embodiments of this specification have been described, those skilled in the art can make additional changes and modifications once they learn the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of this specification.

[0119] Obviously, those skilled in the art can make various changes and deformations to this specification without departing from the spirit and scope of this specification. Thus, if these modifications and deformations of this specification fall within the scope of the claims of this specification and their equivalent technologies, this specification is also intended to include these changes and deformations.

Claims

1. A method for predicting the thickness of a composite plate, characterized in that: The method comprises: Obtaining the assembly thickness data and target hot rolling process parameters of the composite plate to be tested, wherein the assembly thickness data includes sampling point data of the plate head, plate middle and plate tail, and the target hot rolling process parameters are obtained by correlation analysis and screening, including heating temperature, final rolling temperature, total reduction rate and rolling pass reduction distribution; Inputting the assembly thickness data and the target hot rolling process parameters into a prediction model to predict the thickness of the composite plate to obtain a predicted value of the total thickness after hot rolling and cooling, wherein the prediction model is based on the historical assembly thickness data and historical hot rolling process parameters of the composite plate and is obtained through a preset number of trainings; Based on the total thickness prediction value, a composite layer thickness prediction value of the composite plate is calculated.

2. The method according to claim 1, characterized in that The sampling position of the blank assembly thickness data is the first preset range from the width edge of the composite plate, and at least 3 points are collected at the head, middle and tail of each plate number, wherein the blank assembly thickness data includes the first upper composite layer thickness, the first lower composite layer thickness, the first upper composite plate thickness and the first lower composite plate thickness, and the first preset range is 20 mm to 100 mm.

3. The method according to claim 1, characterized in that The specific steps of screening the target hot rolling process parameters by correlation analysis include: Based on the Pearson coefficient, the correlation coefficient between each candidate hot rolling process parameter and the total thickness value actually measured after the hot rolling is completed and cooled is calculated; The candidate hot rolling process parameters whose absolute values ​​of the correlation coefficients are greater than a preset threshold are selected as target hot rolling process parameters.

4. The method according to claim 1, characterized in that: The process of constructing the prediction model includes: Preprocessing the historical batch thickness data and the historical hot rolling process parameters to obtain preprocessing data; Adopting the gradient boosting tree algorithm, taking the preprocessed data as the training set, performing a preset number of iterative training, calculating the sample residual in each iteration, training a new weak learner for fitting, and calculating the weight of the weak learner to obtain a prediction model; The prediction model is optimized using a genetic algorithm to obtain an optimized prediction model.

5. The method according to claim 1, characterized in that The step of calculating the predicted value of the thickness of the composite plate based on the predicted value of the total thickness includes: Determine the relationship coefficient between the total thickness actually measured after hot rolling and the total thickness of the upper and lower layers according to the historical billet thickness data and the historical hot rolling process parameters; Based on the relationship coefficient and the total thickness prediction value, the total thickness prediction value of the upper and lower layers after slitting is calculated; The predicted value of the thickness of the composite layer of the composite board is calculated based on the corresponding relationship between the base layer and the composite layer, and the predicted value of the total thickness of the upper and lower layers after slitting.

6. The method according to claim 1, characterized in that The method further comprises: Comparing the predicted value of the composite layer thickness with a preset tolerance range; When the predicted value of the composite layer thickness is greater than the preset tolerance range, analyzing the historical batch thickness data, the historical hot rolling process parameters, and the current batch thickness data and hot rolling process parameters to obtain deviation factor information; Based on the deviation factor information, the subsequent assembly thickness data or hot rolling process parameters of the composite plate to be tested are adjusted to optimize the prediction accuracy of the composite plate layer thickness.

7. The method according to claim 1, characterized in that Before inputting the batch thickness data and the target hot rolling process parameters into the prediction model, the method further includes: The batch thickness data and target hot rolling process parameters are normalized.

8. A device for predicting the thickness of a composite plate, characterized in that: include: A data acquisition unit, used to acquire the assembly thickness data and target hot rolling process parameters of the composite plate to be tested, wherein the assembly thickness data includes sampling point data of the plate head, plate middle and plate tail, and the target hot rolling process parameters are obtained by correlation analysis and screening, including heating temperature, final rolling temperature, total reduction rate and rolling pass reduction distribution; A model prediction unit, used for inputting the assembly thickness data and the target hot rolling process parameters into a prediction model to predict the thickness of the composite plate, so as to obtain a predicted value of the total thickness after the hot rolling is completed and cooled, wherein the prediction model is based on the historical assembly thickness data and historical hot rolling process parameters of the composite plate, and is obtained through a preset number of trainings; The thickness calculation unit is used to calculate the predicted value of the composite layer thickness of the composite plate based on the predicted value of the total thickness.

9. An electronic device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor is used to implement the steps of the method for predicting the thickness of a composite plate layer as described in any one of claims 1 to 7 when executing the computer program stored in the memory.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for predicting the thickness of a composite plate layer according to any one of claims 1 to 7 is implemented.

Images